Week 5 - Factor analysis and intelligence Flashcards
Factor analysis
A way to demonstrate construct validity
-> Proposed by Cronbach and Meehl (1955) as a way of demonstrating validity. If theoretically related, we should see a significant correlation between constructs.
Items on a test shown to load highly against one or more “factors”
-> Factors are fewer in number than the original variables/items
Two types:
-> exploratory vs confirmatory factor analysis (=EFA vs CFA)
History of factor analysis
Family of statistical techniques used to examine correlations among variables (items or tests)
- > Used in theory validation
- > Used in test construction and validation
- Invented by Pearson in 1901
- Further developed by Spearman in 1904
- Applied by Spearman to cognitive abilities in 1927
Exploratory factor analysis (EFA)
Identifies or finds underlying dimensions/clusters/factors/components in data
Summarises large amount of items or information into fewer scores, based on:
-> correlations among items
-> common variance among items
Also identifies items
-> that are not captured by these clusters
-> and items that are complex- fall equally into two or more different clusters
*neatness is rare, especially in EFA in test construction
Confirmatory factor analysis (CFA)
-> Conducted after EFA with a different sample during test construction
Model-testing function
- Evidence of validity (i.e., structural or factorial validity)
- Tests hypotheses about underlying structure of tests/items
- > i.e., tests if specified relationships / correlations exist in data
- You come up with the model, CFA tells you if it fits the data -> often called ‘goodness of fit’ statistic
Intelligence consists of abilities to
- understand complex ideas
- adapt effectively to the environment
- learn from experience
- engage in various forms of reasoning
- overcome obstacles by taking thought
Interactionalism
Complex concept by which heredity AND environment are presumed to interact and influence the development of intelligence
Major theme in theories of Binet, Wechler, & Piaget.
Factor-analytic theories of intelligence
Focus on identifying the ability or groups of abilities deemed to constitute intelligence.
In fact, that’s why factor analysis was created – to identify commonalities (factors) amongst items or different tests!
Correlations between factors
- Factors produced from items are called 1st order factors
- A correlation matrix of factor scores can also be factor analysed to examine presence of 2nd order factors
- Can also examine for 3rd order factor (global factor- e.g., overall ‘g’ in IQ)
- Rare to go beyond that
US military and intelligence testing
WW1: Army Alpha and Army Beta for verbal ability, numerical ability, ability to follow directions, and knowledge of information, Beta for illiterate unschooled or non-English speaking
WW2: Army General Classification Test and Navy General Classification Test
-> specialised tests, clerical and admin tests and language test also added
